Understanding factors that determine long term outcomes after the onset of major medical conditions is key to identifying effective strategies to improve population health, reduce the prevalence of disability, and eliminate health disparities. Stroke is a leading cause of major disability in elderly Americans. Yet, research into pre- stroke social, contextual, or health characteristics that may reduce the consequences of stroke has been severely hampered by limitations in available data. Most of the preeminent stroke studies are regionally localized and/or initiate data collection upon hospitalization and diagnosis. Few data sources include nationwide samples of stroke survivors followed longitudinally from prior to stroke onset through the long-term recovery period. In data sets collected from only one or a few communities, it is impossible to test the effects of geographic or contextual risk factors on stroke outcomes, or to examine whether racial disparities are modified by such contextual factors. In studies that initiate enrollment at the onset of stroke, examining pre-stroke risk factors typically requires retrospective data collection. Without pre-stroke functional assessments, it is difficult to quantify how much of a functional decrement was induced by the stroke. Although it is difficult and expensive to collect medically verified stroke onset data in a national, prospective cohort, some already existing national longitudinal studies include self-reported stroke. The Health and Retirement Study (HRS) is an ongoing, nationally representative study of Americans aged 50+ that routinely asks participants about history of stroke and other major medical conditions. Since inception, nearly 2,500 incident first strokes have been recorded for HRS participants, but the possibility of misreporting of stroke is a concern when using HRS data to study stroke outcomes. We propose three aims, 1) Quantify measurement error in self-reported incident strokes in the HRS, using linked Medicare claims records as a gold standard;2) Examine national and regional racial disparities in the effects of stroke on mortality, cognitive impairments (assessed with word recall score and the Telephone Interview for Cognitive Status), and physical limitations (instrumental and basic activities of daily living);and 3) Identify pre-stroke characteristics that modify the effects of stroke on health. This research builds on the extraordinary resources in HRS, which includes detailed socioeconomic measures and valuable longitudinal health data. This exploratory project would lay the groundwork for other applications of HRS data, including broader research on determinants of stroke incidence and post-stroke outcomes. These analyses will also provide a model for extensions to other self-reported medical conditions assessed in HRS, such as heart disease and diabetes, so future work could examine both risk factors for onset of these diseases and factors that determine level of functioning in the context of these conditions. PUBLIC HEALTH RELEVANCE: We propose to use data from the Health and Retirement Study, a national longitudinal study of Americans aged 50 and older, to study factors that improve people's functioning after stroke. First, we will compare the stroke records that study participants describe in interviews to their records from Medicare claims. Second, we will look to see if stroke affects black and white patients similarly. Third, we will test whether socioeconomic resources or other characteristics measured before stroke onset reduce the consequences of stroke on survival, memory and cognitive functioning, or physical independence.